If you are asking, "What is Intel® Skylake?," "What is Intel® Purley?," or "What is Intel® Xeon® Scalable processor?," then you have come to the right place.

 And if you are asking, "What is the difference between the three?,” then you have definitely come to the right place!

I have found Intel® is not the easiest company to understand. They seem to have many names for the same things. Having worked with them recently, I realize they do not focus on naming until the very last moment.

The unofficial title to this blog post is “The Dummies Guide to Intel ®Skylake, Intel® Purley and Intel® Xeon ®Scalable Processor.”


Let us start with some simple background information. Intel’s core business is making computer chips, especially the CPUs that do all the calculations in computers. Intel chips are very common in PCs and in datacenter servers. One of Intel’s founders, Gordon Moore, famously predicted that "the number of transistors in a dense integrated circuit doubles approximately every two years,” a prediction now known as Moore's Law. This prediction has, by and large, held true since it was first made in the 1960s, even though it appears to be slowing down in recent years. This continuous exponential improvement of computer power has led us to where we are today with iPhones that are more capable now than supercomputers were 20 years ago. We have Gordon Moore and Intel to thank for that.

There is a lead time for developing such complicated technology and this is where naming of next-generation chips, platforms, and processors starts to become complicated…

What is Intel® Skylake?

Each generation of Intel chips is given an internal "code name” that is used while development is ongoing. The latest generation that is commercially available is code name “Skylake.” More precisely, that is the code name used for the processor's microarchitecture and "Skylake" replaces “Broadwell,” which replaced “Haswell,” which replaced “Ivy Bridge,” which replaced “Sandy Bridge,” which replaced "Nahalem," and so on.

You get the picture…

There's one more wrinkle that needs to be understood in the processor microarchitecture world. Processors do not turn up at the same time for different devices. For example, Skylake has been available in PCs since August 2015; they are already everywhere.

Skylake is only just launching for availability in datacenter servers, which leads us to the next section.

What is Intel® Purley?

Intel® Purley is the server platform that surrounds Skylake. This platform is positioned as the biggest platform advancement since Nehalem in 2009 and it's what is leading to “the end of excuses.” 

To quote Wccftech in the above linked-to article, “Skylake Purley is poised to be the biggest update since the age old Nehalem platform. Along with the improved performance per watt that comes with every article iteration, Skylake EX Purley will actually ship with 6 Channels of DDR4 as opposed to 4. It will also include the AVX 512 instruction set and will boast the 100G OmniPath interconnect. Skylake Purley will also have Cannonlake graphics support not to mention FPGA integration (another important upgrade). The FPGA will be able to execute programmable logic as opposed to the Skylake processor.”

All this leads to the availability of high-performance computing (HPC) for all. Or, in other words, everybody gets their own supercomputer.

You have noticed that everything seems to be changing and devices with artificial intelligence are turning up everywhere—from your car that drives itself to a speaker in your kitchen that replies knowing you better than your wife. In the future, all businesses will need to have machine learning and artificial intelligence to survive. They will all be run like large-scale science experiments; that's why the association with HPC is so apt. This capability to crunch this amount of information this fast has existed only in very specialized computing establishments, such as National Weather bureaus and places such as CERN.

All companies will have this capability the same way all companies now have mobile apps, at least the companies that are still around…

I don't want to overwhelm you but also see 3DXPoint and Non-Volatile Memory Express (NVMe). This technology is the beginning of having memory capabilities far exceeding even that of the human brain. And it never forgets...

So let’s move onto to the final section. I promise to keep it short.

What is the Intel® Xeon® Scalable processor?

Skylake and Purley are examples of code names of microprocessor architectures and server platforms, respectively, not of products. The term "Intel® Xeon® Scalable processors" describes the commercial products you can buy.

Yes, we have finally reached something from Intel that has a product name.

This is Intel's announcement about what it planned to launch on July 11: “Join Intel for its Biggest Data Center Launch in a Decade.”

Fast-forward to the future

And here is what Intel actually launched: "Intel Next-Gen Processors Target Networking, Data Center Workloads."

To quote the article above, "Lisa Spelman, vice president of Intel’s data center group, said the platform can support up to 4.2 times more virtual machines and up to five times more transactions per second than four-year-old systems. The processors also have more storage capabilities, delivering up to five times more operations at up to 70 percent lower latency, she added."

Ericsson has committed to making sure the latest from Intel® is always available in Ericsson Hyperscale Datacenter System 8000, and you can already buy Skylake from us today.

Please see "SDI now with Intel’s latest processor" for more information about Ericsson and Intel® Skylake, Intel® Purley, and the Intel® Xeon® Scalable processor.

Image: Lake Victoria

Prepare your core network Partners & ecosystem Build your cloud and NFV infrastructure

Geoff Hollingworth
Follow Geoff Hollingworth:

Geoff Hollingworth

Geoff is Head of Product Marketing Cloud Systems, responsible for the global positioning, promotion and education of Ericsson’s next generation Cloud infrastructure offerings. He was previously embedded with AT&T in Silicon Valley, leading Ericsson’s innovation efforts towards the AT&T Foundry initiative. He has also held positions as Head of IP Services Strategy for North America and overseeing the Ericsson brand in North America, as well as other roles in software R&D and mobile network deployment. Joining Ericsson more than 20 years ago, Geoff has been based in London, Stockholm, Dallas and Palo Alto. He holds a First Class Honors Bachelors degree in Computing Science and has won the Computing Science Prize of Excellence from Aston University in Birmingham, United Kingdom.